Yes it is, but not all instructions take only one clock cycle - jumps take two, and branches take either one or two.
Presumably this is to do with your WAV player?, you don't want to run the processor at a slow clock speed - certainly at this slow a speed it wouldn't be able to play back your 11KHz samples at anywhere near real time.
You should run the processor as fast as you can, 4MHz should probably be enough, but 20MHz is good if you have a 20MHz processor. You set the playback speed by the time delay between sucessive writes to the D2A, not by crippling the processor.
Unfortunately not true - it would take 2 instruction cycles to output a single sample - any more will take considerably longer. You've got to jump back to the start of the loop, read the data from where ever it's stored, updating the address counter at the same time - loads more than 2 instructions!.
As I mentioned before, it's not common practice to cripple your processor to try and reduce it's speed to match your requirements - you simply write your program to run at the speed you need. Whereas you can quite easily slow your program down (the vast majority of PIC time is spent simply wasting time!), you can't speed it up!.
it would take 2 instruction cycles to output a single sample - any more will take considerably longer. You've got to jump back to the start of the loop, read the data from where ever it's stored, updating the address counter at the same time - loads more than 2 instructions!.